*****  To join INSNA, visit  *****

I have been studying patterns in terrorist and insurgent attacks since 2005.
 In my investigation, the time between attacks almost always follows an
exponential distribution.  The parameter of the distribution is affected by
the geographic boundaries and what constitutes an "attack".  This work has
been replicated many times by others.

What is significant about this finding is the memoryless property, which
means every time a patrol goes on a mission, they have the same probability
of an attack.  The probability of an attack at time *t* is independent of
previous missions.  Predicting a terrorist attack would be like predicting
whether a major league baseball player will hit a home run or that I light
bulb will burn out the next time you turn it on.  This is the same
distribution that governs peoples arrival at a drive thru window at a fast
food place, or customers arriving at the bank.

If we take a closer look at any of these problems, we can find social
explanations that hold more explanatory power.  For example, the
inter-arrival time at the fast food place will be affected by whether it is
lunch time or off-peak time.  In the same way, we must consider social
factors in terrorism.  The alarming thing about the posting, is that just
because you can fit a distribution to something, does not mean you can
predict behavior.


Ian McCulloh, Ph.D.
Major, U.S.Army
Assistant Professor
U.S. Military Academy
West Point, NY 10996

On Sat, Dec 19, 2009 at 1:30 PM, Bienenstock, Elisa <[log in to unmask]> wrote:

> Interesting reading.
> Sent from my iPhone
>  <>

SOCNET is a service of INSNA, the professional association for social
network researchers ( To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.